Devesh Gupta

About Devesh Gupta

I build AI systems that work in production, not just in notebooks.Over the past 3+ years, I\'ve shipped computer vision pipelines, RAG-powered chatbots, and document intelligence platforms across healthcare, construction, and industrial domains. My work has processed images, automated electrical drawing analysis, and reduced manual classification effort by orders of magnitude.Currently at Starco Lighting Inc, I architect end-to-end ML systems, from YOLOv8-based fixture classification to FastAPI microservices with React frontends, serving real business workflows at scale. Before that, at Innova Solutions, I built a RAG medical diagnosis chatbot and a fraud detection system that was recognized at Innova Summit 2023.I hold a Master\'s in Artificial Intelligence from the University at Buffalo (Dec 2025) and have a background from IIT Madras and SRM Institute.What I\'m great at:Production computer vision (YOLOv8, OpenAI Vision API, image classification at scale)LLM application development — RAG pipelines, document intelligence, agentic systemsFull ML lifecycle: data → model → API → deployment (Docker, AWS, Azure)Turning messy monoliths into clean, scalable platformsI\'m currently on F-1 OPT (authorized from Jan 2026) and open to AI/ML Engineer roles where I can build things that matter. Open to opportunities — DM me or reach out at d••••••@buffalo.edu

Experience

1. Ai Engineer

   Starco Lighting Inc

   Jan 2026 — Present · US

   Building production AI infrastructure for the lighting industry — converting manual, document-heavy workflows into intelligent, automated systems.Architected a computer vision pipeline to classify lighting fixture images across 38 categories using YOLOv8 and OpenAI Vision API, reducing manual categorization effort from weeks to hoursTransformed a 2,400-line monolithic Python script into a modular, containerized FastAPI platform with React frontend and Docker deployment — improving maintainability and enabling team-wide usageBuilt a document intelligence system that parses electrical drawings (PDFs) and extracts structured data — automating fixture identification and cross-referencing for lighting auditsDeveloped automated product cross-referencing pipelines to match legacy fixtures against current catalogs, directly supporting sales and procurement teamsDesigned and deployed REST APIs enabling non-technical stakeholders to interact with ML models through clean interfaces

Education

• Indian Institute of Technology, Madras

  Diploma of Education, Data science and programming

• University at Buffalo

  Masters, Artificial Intelligence

• SRM IST Chennai

  Bachelor's degree, Mechanical Engineering

• Indian Institute of Technology, Madras

  Associate of Science - AS, Data science and programming